Anzeige
Erschienen in:
10.01.2022 | Original Article
Urinary metabolomics to develop predictors for pediatric acute kidney injury
Erschienen in: Pediatric Nephrology | Ausgabe 9/2022
Einloggen, um Zugang zu erhaltenAbstract
Background
Acute kidney injury (AKI) is characterized by an abrupt decline in glomerular filtration rate (GFR). We sought to identify separate early urinary metabolomic signatures at AKI onset (with-AKI) and prior to onset of functional impairment (pre-AKI).
Methods
Pre-AKI (n=15), AKI (n=22), and respective controls (n=30) from two prospective PICU cohort studies provided urine samples which were analyzed by GC-MS and DI-MS mass spectrometry (193 metabolites). The cohort (n=58) was 8.7±6.4 years old and 66% male. AKI patients had longer PICU stays, higher PRISM scores, vasopressors requirement, and respiratory diagnosis and less commonly had trauma or post-operative diagnosis. Urine was collected within 2–3 days after admission and daily until day 5 or 14.
Results
The metabolite classifiers for pre-AKI samples (1.5±1.1 days prior to AKI onset) had a cross-validated area under receiver operator curve (AUC)=0.93 (95%CI 0.85–1.0); with-AKI samples had an AUC=0.94 (95%CI 0.87–1.0). A parsimonious pre-AKI classifier with 13 metabolites was similarly robust (AUC=0.96, 95%CI 0.89–1.0). Both classifiers were similar and showed modest correlation of high-ranking metabolites (tau=0.47, p<0.001).
Conclusions
This exploratory study demonstrates the potential of a urine metabolite classifier to detect AKI-risk in pediatric populations earlier than the current standard of diagnosis with the need for external validation.
Graphical abstract
×
![]()